The Future of the Localization Industry - Lionbridge
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Focus Focus: AI and Making Localization Tech Work for You MultiLingual Media LLC • 319 North First Avenue, Suite 2 • Sandpoint, Idaho 83864-1495 USA The Future of the Localization Industry Kajetan Malinowski Jaime Punishill Kajetan Malinowski is senior director of Proposition Jaime Punishill is chief marketing officer at Management for Lionbridge, and is responsible for Lionbridge, and is responsible for leading global Lionbridge’s product strategy and for developing an marketing and proposition development efforts. AI-enabled localization platform. Malinowski holds an He earned a bachelor of arts in history and political MBA from UQAM /Warsaw School of Economics. science from Stanford. January/February 2021 1
Fo c u s What is the future of the localization industry? To understand where we’re headed, it is helpful to consider how we got here. When I started working in localization 14 years ago, things were a lot different. We worked on multi-month, behemoth projects. Success was measured by how “good” the translation was, as defined by its linguistic quality and technical compliance. Localization was totally divorced from business outcomes. It was like the old medical joke, “the operation was successful, but the patient died.” Language service providers (LSPs) were treated like an extension of internal localization departments — a factory where the content was “processed.” Many business functions that are now critical started this way. For example, the early days of digital marketing were quite similar. Relegated to the sidelines, digital marketing teams were separate from the main marketing department. These “non-critical” functions tended to be underfunded and understaffed. They were not viewed as core or strategic to the business. Time, of course, proved the old-line marketers wrong. Digital wasn’t a side show — it was to evolve into the prime acquisition channel for companies. Fast-forward to 2021, and we see that the pandemic has erased whatever resist- ance or sluggishness was left in companies towards digital channels. Digital transformation is now front and center as we’re living in a digital-first universe. The shift has forced companies to reevaluate their business practices and finally prioritize digital marketing strategies. The rise of the multimarket In 2021, even more businesses will move to a digi- digital experience tal-first approach. This shift will bring exciting changes As the shutdowns that cascaded across the globe in that have profound implications for the localization in- 2020 proved, many companies were not ready for the dig- dustry. As a result, the localization industry will be able to ital-first universe. About ~10% of a brand’s digital content take its rightful place at the strategic table as a key com- gets localized, and even fewer global companies provide ponent of every company’s multichannel experience. Both a truly multilingual support experience across the full companies and LSPs that embrace this shift will gain a spectrum of their customer support channels, according competitive edge over slower-moving competitors, ush- to CSA Research’s “Localization Depth and Language ering in a new age of digital experience and accelerating Choice.” international growth. The contrast between the physical and online worlds is stark. In person, you can operate in your native language Turning the corner: An explosion of and you are less aware of what is available in another phys- content and the dawn of AI ical location. Online shoppers can see exactly what’s avail- While digital tactics, processes and best practices are able but are often faced with content that is not in their slowly taking over, the digital world is getting infinitely native language. Language remains the ultimate frontier of more complicated. Video. Mobile. Social. User-generated the frictionless customer journey. content. Email. SMS. Augmented reality. Along with the Leaders know what others have just discovered: to be explosion of channels and media, an exponential increase effective, localization has to stop being just about words, in velocity and volumes stresses the system. There are no or a mere afterthought. It must incorporate meaning, more 12-month campaigns and launches. Years-long in- comprehension and emotion, the key drivers that moti- itiatives turned to months to weeks and now, mere days. vate consumers to act. The amount of content generated every year is growing Innovative digital teams are more attuned to the rising exponentially. Nearly 90% of the world’s data was created significance of language, cultural adaptation, and the tar- in the 24 months prior to a 2018 Forbes study — no doubt geting of market-specific buying patterns — but they still the last 24 months have seen similar meteoric growth. struggle with an internal culture and process that focuses Generating fresh, relevant content is the lifeblood of de- on the physical world first and the digital world second. mand generation; it increasingly separates digital winners Many marketing teams still undervalue internationaliza- and losers. Leaders realize that to sustain a constant, ev- tion, and operate inefficient content supply chains that op- er-growing flow of content, they must change the way they timize for their host country or language, and not for the create, and localize that content to meet the needs of the content’s use in multiple markets simultaneously. Couple buyers in their target markets. that with the continued focus on process and linguistic Localization teams are responding. First, there was an purity by most localization teams, and you realize that the emphasis on transactional efficiency, connectivity and go-to-market and localization teams can’t be further apart. the simplicity of the localization technology stack. This 2 January/February 2021
Focus has given rise to next-gen computer-assisted translation Change #1: Augmented (CAT) and translation management system (TMS) tools. translation arrives Today, companies increasingly expect LSPs to integrate In the years to come, digital experiences will displace and automate localization processes as part of the local- in-store experiences evermore. As a result, their current ization service. relationship will become inverted, with in-store experi- Transactional speed and efficiency are just parts of the ences relegated to supporting digital experiences going problem. With growing demands, greater volumes and forward. This shift started long before the pandemic, but shorter turnaround times, translation cost-cutting meas- it’s the post-COVID reality that will normalize it. Ap- ures are the next tactic teams turn to. The long-awaited ple understood that dynamic and completely redesigned potential of machine translation (MT) provided a new their retail experiences, turning them into digital-pow- glimmer of hope in 2017 with the introduction of neural ered showroom experiences. That is the path that many machine translation. Neural MT’s quantum leap in per- brands will follow in the digital-first, multichannel world. formance should unlock the next wave of efficiency while Knowing when to deliver content, in which language and increasing the quality of MT for many more languages, through which channel will be an essential part of the mul- making it applicable to use cases that were not previously timarket, competitive equation for 2021 and beyond. considered. Substantial advancements in the use of neu- MT was the first use of AI in our industry, and AI is ral networks and artificial intelligence drove this break- about to drastically transform the way translators operate through, and both rely on large quantities of high-quali- in the years to come. Their focus will gradually shift from ty language data–a new focus for innovative localization translating source content to correcting and augmenting teams. machine-generated translations. They will add relevant The 25-year journey of the localization industry has context and fill in language or market-specific gaps in the seen amazing transformations and progress. We are just translated narrative. There will also be other applications starting to unlock the real value creation in this industry. of AI that further improve translator efficiency. These ap- The next stage of evolution will capitalize on insights and plications include type-ahead functions, do-not-translate analysis about language, tone, sentiment, readability, ac- blockers, inline readability highlights, the detection of bi- cessibility, and buyer journey impact. In our work we see ased or offensive terms, and the autocorrection of words signs of transformation coming from early adopters. that will ensure adherence to terminology glossaries or Figure 1: Volume and quality drive different quality levels and production processes. Source: Lionbridge. January/February 2021 3
Fo c u s brand style guides. It won’t be long before natural-lan- a pedantic universe that only cares about perfect transla- guage content generation — powered by language pre- tions, despite reams of evidence that most users will accept diction models like Generative Pre-trained Transformer 3 many less-than-perfect translations just to get the 90+% of (GPT3) — is incorporated into the translation process as content never localized into a zone they can understand. well. There are two main governors creating the content gap- Most industry discussion about the impact of AI focus- cost, and the persistent belief that only perfect quality will es on MT and translators, but we believe AI’s influence do. Soon, innovators won’t choose which content is going will be far greater on the broader localization process. AI to be localized. Instead, they will target 100% localization will unleash an explosion of content to be localized and for all markets and optimize their budgets by choosing generate a significant growth in the marketplace. different quality and service levels based on the necessary demands of the content by use case. Not every piece of Change #2: Augmenting localization content needs white glove treatment. Well-trained and end-to-end managed MT is proven to handle vast amounts of con- The notion that localization teams, and their LSP part- tent from the bottom of the content pyramid. Adding on ners, will morph into global content production teams increasing amounts of human attention and experience is not new — CSA Research mentioned it, for example, takes you from the bottom of the pyramid to the top. in June 2020’s “The Future of Language Services.” With- Together, the AI-driven content journey and an under- out scalable AI, it wasn’t possible, but that future has standing of “good enough” quality needs will enable firms arrived. Taking advantage of these new capabilities re- to quickly shift from localizing 10% of their digital content quires businesses to rethink how content or experiences to localizing 90% of it. This will bring a frictionless path to are produced. Naturally, a focus on internationalization purchase, unlock forgone revenue and improve their cus- best practices is central to this strategy. Many localization tomer’s experience in all channels teams still operate in organizations that haven’t optimized their people, process, and technology to the multimarket Change #4: Marketing and localization world they operate in. The digital-first universe that was functions will merge thrust upon organizations in 2020 makes this deficiency We expect more and more marketing organizations to even more pronounced. How can AI be harnessed beyond take on the localization remit from product and devel- translation? opment teams. Why? Increasingly, marketing teams are In agile software development, there is a notion called responsible for brand, growth, customer experience, on- “shift left.” Many things done late in the development cy- boarding, and digital transformation. None of those things cle can be done much earlier, adding speed and efficiency can be done successfully for a multimarket organization while improving quality. For example, a coder will write without mastering the multilingual, multicultural, multi- both her code and test cases at the same time, a process channel, and multi-buyer journey differences that exist in that used to be done by two people at separate points in the today’s world. process. In localization workflows, the shift left has started. Marketing teams are more customer-centric, more Take many of the quality processes created to ensure good data-driven, and more agile than other business units be- translation, executed after the translator returns the work. cause they have had to be these things to survive digital What if those same quality measures could be applied to transformation. This change is already starting in some of the source content before the content is translated? What if the organizations that have been leaders in localization, the search engine optimization (SEO) terms often applied and we expect many other companies to follow suit. As after the fact, were instead maintained in a data reposito- new content journeys emerge and reimagined definitions ry and applied to the translation process in flight? Both of of quality take hold, the logic of creating one single glob- those examples are real and in limited use today, and both al content and customer experience team to support the are enabled by the broad category of AI disciplines. multimarket strategy of globally minded organizations will become obvious to many. Change #3: Unleashing a torrent of “good enough” content The multimarket content journey for The most important question facing large scale retail- today’s multichannel experience ers, technology companies, travel and hospitality brands As marketing and localization teams come together to is, “What is good enough quality?” as measured by end-us- reimagine how they create content and experiences, they er acceptance. Many localization professionals are stuck in will rethink their process from end-to-end, incorporating 4 January/February 2021
Focus these new AI-driven insights at each stage of the process. CMS or deposited in the digital asset management system, What is needed is a construct that connects all parts of it will be auto-ingested by the localization platform of the the content cycle and is engineered to drive and support organization or their LSP. This is more than just a work- business outcomes (see Figure 2). flow automation; leaders will recognize the opportunity to The first two phases — content planning and content tag all content so that they can identify the content for do- authoring — are closely related to the last phase, content main, language, persona, readability, offensive terms and performance. These three phases are centered around inclusive language, brand fit, and other dimensions. This marketing optimization activities and wrap around the sets the stage for the best way to transform the content ones focused on localization. Each phase is important and into the target state. constitutes an essential part of the journey: 4. Content conversion. In recent years, content types 1. Content planning. Centered around the planning of have multiplied, increasing process complexity. There content marketing activities and the budgeting of those are limits to what can be done with content in its origi- activities, this stage is where organizations set their con- nal form, necessitating its conversion to text for further tent strategy and indicative budgets for creation and local- transformation and enrichment. Every content conversion ization, using a matrix of buyer profiles, market dynamics, poses risks of data loss or formatting breakage. This phase and growth objectives. ensures content fidelity during the extraction process. 2. Content authoring. During this phase, companies 5. Content transformation. In the old localization mod- craft content that aligns with their content strategy goals. el, this is where most of the effort, and hopefully value, was They are aided by insights into where they have content expended and created. But as you can see, the new content gaps relative to what buyers are looking for, how to best journey uses preprocessing and content analysis to make author that content to have the greatest discoverability in this step purely about the translation or transcreation of search engines and tagged to best support personalization the content into multiple languages and its adaptation to and recommendation tools. Altogether, these things max- different locales. Content can be assessed and targeted for imize customer engagement and reduce friction on the different quality levels and process flows can be matched paths to purchase. by need. Content will be matched with the best-fit linguists 3. Content ingestion. Once content is authored in the and other transformation agents. And with improved MT, Figure 2: Most important stages of content life cycle. January/February 2021 5
Fo c u s better integration of MT and productivity enhancements consumers. By working in tandem, marketing and lo- such as type ahead, the efficiency of linguists will improve calization teams can harness the power of AI-driv- as well. en insights to make informed choices and answer 6. Content enhancement. More video, audio, and in- key questions such as: teractive modules are being created than ever before, and the content enhancement stage reflects the rapid increase • Which content should be localized? in non-word work. This includes things like formatting • How can a piece of content be improved to make it and layout changes to adapt content to local requirements more localizable and impactful? and specificities. Of course, this stage will encompass ad- • Which content is easier to read for which audience? ditional quality — and other linguistic analysis steps — to • What content will have the most impact? ensure it’s on brand, on point, at quality, on target and fit • How can one assure that content is on brand, in voice, for purpose. Because so much analysis has been shifted to isn’t offensive, and is appropriate for the market in the authoring stage, this new linguistic quality stage will question? be more automated, with exceptions made only for addi- tional human review. AI content intelligence helps combine SEO, language 7. Content delivery. Once the content has passed all and content research into one seamless process that em- the gates at the last stage, it will be auto routed back to powers demand gen and engagement teams to craft content the appropriate client system. At this stage, content gets that is insightful, that resonates with their readers and that published on the target medium directly via the CMS or translates well. Pre-analysis will lead to better-informed digital asset management system. decisions about whether the content should be localized, 8. Content performance. This is one of the most im- de-biased, edited or rewritten for a market to maximize portant steps. It’s at this stage that companies assess the impact for those buyers and their unique path to purchase. performance of their content by measuring key perfor- Whether digital or analog, content has always been es- mance indicators, such as conversion rates, visits and en- sential to brand, customer engagement, persuasion, and gagement. The key enabler to this was the addition of tag- conversion. An AI-powered content workflow will amplify ging early in the process. Content performance closes the content. Closing the yawning gap between the quantity of whole cycle and feeds into the first step of the cycle, con- source and target market content, and the connection of tent planning, while providing valuable information. Im- localization work to measurable outcomes, will elevate lo- portantly, tracking and assessing whether content should calization to the strategic table. AI will enable localization have been localized at all will reshape how multimarket teams to attack both of those issues in concert with their content strategy is conceived and executed. partners in marketing and customer experience. A fully implemented content journey approach will There has never been a more exciting time in the local- deliver valuable insights to brands that are seeking to es- ization business. No longer an afterthought, today, locali- tablish engagement and a lasting relationship with their zation is front and center. Share your thoughts on the future of localization with a three-minute poll on multilingual.com/poll Lionbridge partners with brands to break barriers and build bridges all over the world. For more than 20 years, we have helped companies connect with global customers and employees by delivering localization and training data services in 350+ languag- es. Through our world-class platform, we orchestrate a network of one million passionate experts in 5000+ cities, who partner with brands to create culturally rich experiences. To learn more visit www.Lionbridge.com Reprinted with Permission • MultiLingual • Volume 32 Issue 1 • January/February 2021 © 2021 MultiLingual Media LLC, 319 North First Avenue, Suite 2, Sandpoint, Idaho 8364-1495 USA 6 January/February 2021
You can also read